Floods are one of the most common types of natural disasters. They annually affect vast amounts of people and cause severe economic losses. While fluvial, coastal, and flash floods are well studied, pluvial floods (rain related) have received modest attention from researchers and decision-makers in comparison. There are several reasons for this, one is that it has been considered a fixed problem with infrastructure and other engineered solutions and another is that they are generally undramatic and small-scale. However, as cities expand, the environment’s ability to retain and dispose of excess water is inhibited and as the frequency of extreme weather events is expected to increase due to climate change, the risk associated with pluvial floods has become increasingly recognized. Commercial and open-source Urban pluvial flood models tend to require advanced modelling expertise, considerable computational power, large amounts of input data and are often expensive. Consequently, there is less knowledge of flood inundation caused by pluvial floods compared to other types. This thesis investigates the Bluespot model, which aims to provide an approachable tool to generate an overview of the effects of pluvial floods in urban areas. The model requires few input data and is relatively simple to perform. Results from the model are compared to the August 2021 flood event in Gävle, Sweden.The study finds that results ranged from accurate to over- and underestimated. Slope and incoming water were found to affect the outcome most. Blue spots without the influence of streams or other waterways, with a distinct slope were mapped with accuracy and showed consistency with coarser resolutions. Consequently, underpasses in the road network were mapped with especially good consistency. Further, blue spots within close distance to large flow accumulation were underestimated and the accuracy tended to decrease with a coarser resolution. The model cannot account for water outside blue spots, thus, when large volumes of water accumulate and spread beyond these boarders it generates poor results. These areas were found to be efficiently indicated by generating a heatmap from high-flow accumulation points. Thus, indicating low confidence and where a hydraulic flood model should be performed. Depending on the scope a 1-3m resolution is recommended for investigating effects on property etc and a 5-10m resolution is sufficient for investigating underpasses, however, a finer resolution will generate more accurate results.

Urban pluvial floods occur during extreme rain events and both occurrence and magnitude of these floods are expected to increase. Preserving or constructing green areas in urban areas has been shown to mitigate and control these floods. The common way to evaluate flood risks is to set up a rainfall-runoff model, but these studies are often case related and only investigate the soil characteristics for that specific case. Multiple studies have also stated that the difference between major and minor flooding effects is connected to the antecedent soil moisture content. This thesis attempts to investigate how different soil characteristics influencing infiltration affect the hydraulic response in two Swedish urban catchments and if antecedent soil moisture is a critical factor. To evaluate the hydraulic response, a two-dimensional surface runoff model of two different urban catchments was forced with a hyetograph of a CDS-rain with a return period of 100 years. The simulations were conducted with three different soil types for all urban green areas: clay, sandy loam and sand, and three different antecedent soil moisture contents for clay and sandy loam. Flood extent and discharge from catchment area was evaluated, as was flood depth and overland flow in 16 chosen evaluation points. The results showed that with decreasing infiltration rate of a soil and with increasing antecedent soil moisture content, the severity of the flood and discharge at the catchment outlet was increased. It was also concluded that soil type affects flood extent, flood depth, overland flow and discharge from catchment to a greater extent than antecedent soil moisture.

Les contours théoriques en hydrologie, hydraulique et les outils de calcul correspondants sont largement développés et utilisés dans le monde. Cependant en parallèle, des problématiques importantes surviennent pendant les crises sans pouvoir être résolues et des solutions développées peinent à être implémentées. En plus, le changement climatique ne va pas faciliter les choses. Pour noircir le tableau, les moyens économiques locaux en France ne vont pas augmenter pour aider les communes à s’attaquer au problème. L’objectif de cette thèse est de conduire une investigation des moyens à disposition pour améliorer notre connaissance locale des concepts en lien avec le pluvial pour permettre une modélisation efficiente. La méthodologie proposée est composée de trois étapes évolutives incluant : 1. Une analyse approfondie des données topographiques locales ; 2. L’évaluation des interactions entre les écoulements de surface et le souterrain ; 3.Une approche intégrée permettant de modéliser les inondations générées par la pluie en zone urbanisée. Les résultats de l’étape 1 montrent que la donnée topographique est essentielle pour la définition des chemins d’écoulement et impactent significativement les résultats de modélisation hydrauliques. Ceci conduit à l’étape 2 lors de laquelle on observe que les débordements provenant du réseau souterrain contribuent à l’inondation mais seulement en partie. Les volumes d’inondations générés par le ruissellement de surface devraient être inclus dans les modèles d’inondation. L’étape 3 présente une configuration possible de modèle intégré permettant de mieux représenter les processus réels en jeu. / Theoretical background about hydrology, hydraulics and computational tools and methods are widely developed and worldwide used. In the same time however, important issues during flood crisis are not solved and practical solutions take time to be implemented. On top of that, ongoing climatic change will not make things easier and intense events will increase in frequency. To worsen the picture, local economic means in France will not increase to help municipalities and local communities to tackle the issue. The objective of this thesis is to investigate on the available ways to improve our local knowledge of stormwater related concepts to allow an efficient modelling. The proposed methodology consists in a three-step-approach including: 1. A thorough analysis of local topography data; 2. The assessment of sewer-surface interactions; 3. An integrated approach to model pluvial flood in urban areas. The results of Step 1 show that topography data is essential in flow path definition and significantly impacts hydraulic modelling results. This leads to Step 2 where it is seen that sewer overflow is one aspect of urban flood issues but represents only part of flood sources. Overland flow generated by runoff should be included in flood models. Then Step 3 presents that integrated urban pluvial modelling is possible with existing tools and can represent the real processes better. This proposed modelling approach should not be disconnected from the reality of stormwater management practical aspects and current constraints. It is shown how complementary actions can be taken to enrich local knowledge and memory thus allowing a more efficient and wiser modelling process.